flywheel

Flywheel

A Claude Code plugin that encodes April Dunford's positioning and sales pitch frameworks, Madhavan Ramanujam's monetization framework, and Matt Lerner's growth framework into structured workflows. The friction is the product — each skill forces the specificity that generic AI marketing skips.

Outputs are plain markdown, git-tracked, and compound over time. Each session's decisions feed the next.

Skills

/fw:position — Positioning Canvas

Works through Dunford's 5-component positioning sequence:

1. Competitive alternatives — what customers do if you don't exist
2. Unique attributes — what you have that alternatives genuinely lack
3. Value — attribute → capability → outcome chains
4. Customer segments — who cares most, specifically
5. Market frame — the category that makes your value obvious

Produces a positioning canvas at docs/positioning/current.md that all other skills read automatically.

/fw:copy — Copy Artifacts with Drift Detection

Translates positioning into specific copy artifacts:

• Landing page, pitch, bio, outreach, talk abstract, case study, one-liner

The key feature is the drift detector — it traces every value claim in the draft back to the positioning canvas. Anything that can't be traced gets flagged as a generic insertion. You decide: remove it, rewrite it, or acknowledge the departure.

For short-form artifacts (one-liners, subject lines), generates 5 variations with likelihood ratings instead of a single draft.

Saves to docs/copy-tests/ with outcome notes for recording real-world results.

/fw:pitch — Sales Pitch Storyboard

Encodes April Dunford's sales pitch framework from her book Sales Pitch. Builds a 7-step storyboard for a live sales conversation or pitch deck:

1. The Insight — your POV on the market that reframes buyer thinking
2. The Old Game — how buyers currently approach the problem (fair, not strawmanned)
3. The New Game — the new buying criteria the insight implies
4. The Perfect Solution — what an ideal product in the new game would look like (abstract)
5. Your Differentiated Value — your product as the closest thing to the ideal
6. Proof — specific evidence for every value claim
7. The Ask — the proportional next step

Reads the positioning canvas automatically and runs drift detection on value and proof claims — nothing in the storyboard should be ungrounded. Supports revision mode for updating specific sections after a pitch falls flat, and deck mode for slide-by-slide notes.

Saves to docs/pitch-storyboards/ with outcome notes so future sessions know which insights and asks have actually landed.

/fw:monetize — Pricing & Monetization Strategy

Encodes Madhavan Ramanujam's framework from his book Monetizing Innovation. Produces a defensible pricing decision in 6 steps:

1. Leaky Bucket Analysis — classify every unique attribute as differentiator, filler, loser, or leader (diagnoses feature shock and hidden gems)
2. WTP Signals — surface every willingness-to-pay signal with its source; flag gaps as next-action interview questions
3. Segmentation by WTP — regroup customers by what they'd pay, not demographics, with a price corridor per segment
4. Product Configuration — good/better/best tiers where each tier serves a WTP segment and anchors on a differentiator
5. Monetization Model — subscription, usage, outcome, freemium, per-seat — chosen to match the buyer's mental model
6. Price Points & Corridor — list, floor, and ceiling per tier, every number cited to a signal or competitive reference

Supports a WTP interview guide mode (/fw:monetize wtp) that produces a structured document the founder takes to real customer conversations — Van Westendorp questions, feature value probes, and competitive price probes, all grounded in the positioning canvas.

Strict drift detection on price points: any number without a cited signal or reference is flagged as the Ramanujam failure mode (guessing at price). Saves to docs/pricing/current.md as a singleton decision that evolves, with archived versions and WTP interview findings alongside.

/fw:grow — Growth Experiments

Designs experiments using Lerner's behavior-first framework:

1. Name the behavior — a specific person doing a specific action, not a metric
2. Diagnose the barrier — awareness, comprehension, trust, friction, or motivation
3. Form the hypothesis — the smallest intervention that tests the barrier
4. Design the experiment — who, what, how long, how measured (must be under 1 week)
5. Define success criteria — success, failure, and ambiguous thresholds before running

Produces a handoff-ready experiment card at docs/growth-experiments/. If someone who wasn't in the session can't run the experiment from the card, it's not done.